Data Annotator (Scholar) — Nidra image output editing dataset annotation
Annotated AI training datasets for an image output editing project by comparing input and output images to identify visual differences. Labeled regions and attributes associated with subject pose, object changes, background edits, and style transformations using structured project guidelines. Applied quality controls to meet SLA requirements and stay within an acceptable error threshold while producing transformation instructions. • Labeled visual change regions using annotation tools. • Wrote action-oriented transformation instructions to guide model training. • Skipped or escalated tasks involving NSFW content, identical images, or unrelated pairs. • Performed peer reviews to ensure labeling accuracy and consistency across the dataset.